Personalised Accelerometer Cut-point Prediction for Older Adults' Movement Behaviours using a Machine Learning approach

Nonso Nnamoko, Luis Adrián Cabrera-Diego, Daniel Campbell, George Sanders, Stuart Fairclough, Ioannis Korkontzelos

Abstract: Highlights•A model is developed to predict physical activity cut-points on accelerometer based on individual characteristics•Post data collection analytical process helps towards a standardised method for characterising physical activity•Multiple features calculated from raw accelerometer data was used to enrich the feature set for training machine learning•Personalisation was achieved by combining raw accelerometer data with person-specific data e.g., blood pressure